MPRAsnakeflow experiment QC report

Good Spearman correlation across replicates!

The data has a good Spearman correlations across replicates using a barcode threshold per oligo of 10! Spearman correlation of DNA is 1.00, RNA is 0.98 and DNA/RNA ratio is 0.96.

Overall quality metrics

Table explanation

median rna read count: Median of RNA read count for oligos that passed filtering to determine sufficient coverage in terms of read count. Value is the median of all replicates.
median barcodes passing filtering: Median number of barcodes across tested sequences that passed filtering to determine if there was sufficient barcode to oligo coverage. Value is the median of all replicates.
pearson correlation: The correlation of log2 RNA/DNA ratios across tested sequences as a measure of replicable activity signal. Value is the median of replicate comparisons using only oligos with >= 10 barcodes.
fraction oligos passing: Fraction of tested sequences that passed filtering of the mappable sequences to determine if the designed library was sufficiently recovered. Value is the median of all replicates and using only oligos with >= 10 barcodes.

median barcodes passing filtering	median rna read count	pearson correlation	fraction oligos passing
199	13676	0.97	0.98

DNA over RNA counts

Plotting normalized counts of DNA vs RNA (median across replicates). Only oligos within all replicates are shown. We should see a variation within the RNA count data (along the y axis). If count data between RNA and DNA is highly correlated (e.g. follows the identity line) there is no variation between designed oligos. This is an indication that RNA is inflated with DNA and the DNA digestion before creating cDNA did not work as expected.

Oligo correlation

Oligo correlation plots of DNA, RNA and DNA/RNA ratios across replicates. First tab shows plots using (in average) 2395 oligos with a minimum number of 10 barcodes. Second tab shows all 2438 oligos that have assigned barcodes.

Condition	A	B	#Oligos A	#Oligos B	#Oligos Joined	DNA spearman	RNA spearman	Ratio spearman	DNA log2 pearson	RNA log2 pearson	Ratio log2 pearson
HEPG2	1	2	2395	2397	2395	1.00	0.98	0.96	1.00	0.98	0.96
HEPG2	1	3	2395	2396	2395	1.00	0.98	0.97	1.00	0.98	0.97
HEPG2	2	3	2397	2396	2396	1.00	1.00	0.99	1.00	1.00	0.99

Condition	A	B	#Oligos A	#Oligos B	#Oligos Joined	DNA spearman	RNA spearman	Ratio spearman	DNA log2 pearson	RNA log2 pearson	Ratio log2 pearson
HEPG2	1	2	2438	2438	2438	0.99	0.98	0.96	0.99	0.98	0.96
HEPG2	1	3	2438	2438	2438	1.00	0.98	0.97	1.00	0.98	0.97
HEPG2	2	3	2438	2438	2438	1.00	1.00	0.99	0.99	1.00	0.99

Experiment statistic

The total number of oligos in this experiment is 2439 (defined by the assignment) with 1144171 associated barcodes.

In average across replicates we see 2438 from 527485 average barcodes in the count data and around 308720 barcodes where not in the assignment.

condition	replicate	oligos dna/rna	matched barcodes	unknown barcodes dna/rna	% matched barcodes	total dna counts	total rna counts	avg dna counts per bc	avg rna counts per bc	avg dna/rna barcodes per oligo
HEPG2	1	2438	526817	307782	63.12	21004709	55170350	25.17	66.10	216.09
HEPG2	2	2438	523270	227374	69.71	20013767	36061744	26.66	48.04	214.63
HEPG2	3	2438	532367	391005	57.65	29614277	55689751	32.07	60.31	218.36

Experiment	Barcodes	Counts	Average counts	Assigned barcodes	Assigned counts	Average assigned counts	Fraction assigned barcodes	Fraction assigned counts
HEPG2.1.DNA	1712563	21963700	12.83	541735	19392468	35.80	0.32	0.88
HEPG2.2.DNA	1643592	20988551	12.77	539266	18561968	34.42	0.33	0.88
HEPG2.3.DNA	2112322	30931193	14.64	551785	27349232	49.57	0.26	0.88
HEPG2.1.RNA	2779664	57511432	20.69	564071	50282968	89.14	0.20	0.87
HEPG2.2.RNA	2199226	37761972	17.17	557727	33004056	59.18	0.25	0.87
HEPG2.3.RNA	2777028	57875580	20.84	564813	50680776	89.73	0.20	0.88

Histograms barcodes per oligo, counts per barcode

Histogramm of number of barcodes per oligo and the number of counts per barcode devidied by DNA and RNA. Median is red, mean is blue.

Activity

Violin and box plots of the log2 fold change of all oligos by the assay. Grouped by labels if set, otherwise NA. First tab shows plots using (in average) 2395 oligos with a minimum number of 10 barcodes. Second tab shows all 2438 oligos that have assigned barcodes.

--- title: "MPRAsnakeflow experiment QC report" page-layout: full format: html: code-fold: true code-tools: true embed-resources: true anchor-sections: true smooth-scroll: true toc: true theme: light: flatly dark: darkly jupyter: python3 --- ```{python} #| echo: false # Imports from IPython.display import display, Image, Markdown import pandas as pd import json ``` ```{python} #| tags: [parameters] #| echo: false ``` ```{python} #| echo: false oligo_cor_merged_table_thresh = pd.read_csv(f"{workdir}/{statistics_all_oligo_cor_thresh}", sep='\t') oligo_cor_merged_table_all = pd.read_csv(f"{workdir}/{statistics_all_oligo_cor_all}", sep='\t') dna_spear = oligo_cor_merged_table_thresh['DNA_spearman'].min() rna_spear = oligo_cor_merged_table_thresh['RNA_spearman'].min() ratio_spear = oligo_cor_merged_table_thresh['Ratio_spearman'].min() ``` ```{python} #| echo: false threshold_dna_spear = 0.85 threshold_rna_spear = 0.95 threshold_ratio_spear = 0.75 important = dna_spear < threshold_dna_spear and rna_spear < threshold_rna_spear and ratio_spear < threshold_ratio_spear warning = dna_spear < threshold_dna_spear or rna_spear < threshold_rna_spear or ratio_spear < threshold_ratio_spear if important: display(Markdown(f''' ::: {"{.callout-important}"} ## Low correlation! Spearman correlation across replicates is low using a barcode threshold per oligo of {thresh}! Spearman correlation of DNA is {dna_spear:.2f}, RNA is {rna_spear:.2f} and DNA/RNA ratio is {ratio_spear:.2f}. Minimum allowed values are {threshold_dna_spear:.2f}, {threshold_rna_spear:.2f}, and {threshold_ratio_spear:.2f}. ::: ''')) elif warning: display(Markdown(f''' ::: {"{.callout-warning}"} ## Low correlation warning! At least one (not all) Spearman correlations across replicates is low using a barcode threshold per oligo of {thresh}! Spearman correlation of DNA is {dna_spear:.2f}, RNA is {rna_spear:.2f} and DNA/RNA ratio is {ratio_spear:.2f}. Minimum allowed values are {threshold_dna_spear:.2f}, {threshold_rna_spear:.2f}, and {threshold_ratio_spear:.2f}. ::: ''')) else: display(Markdown(f''' ::: {"{.callout-tip}"} ## Good Spearman correlation across replicates! The data has a good Spearman correlations across replicates using a barcode threshold per oligo of {thresh}! Spearman correlation of DNA is {dna_spear:.2f}, RNA is {rna_spear:.2f} and DNA/RNA ratio is {ratio_spear:.2f}. ::: ''')) ``` # Overall quality metrics <details> <summary>Table explanation</summary> - **median rna read count:** Median of RNA read count for oligos that passed filtering to determine sufficient coverage in terms of read count. Value is the median of all replicates. - **median barcodes passing filtering:** Median number of barcodes across tested sequences that passed filtering to determine if there was sufficient barcode to oligo coverage. Value is the median of all replicates. - **pearson correlation:** The correlation of log2 RNA/DNA ratios across tested sequences as a measure of replicable activity signal. Value is the median of replicate comparisons using only oligos with >= `{python} thresh` barcodes. - **fraction oligos passing:** Fraction of tested sequences that passed filtering of the mappable sequences to determine if the designed library was sufficiently recovered. Value is the median of all replicates and using only oligos with >= `{python} thresh` barcodes. </details> ```{python} #| echo: false with open(f"{workdir}/{qc_metrics_file}", 'r') as file: qc_data = json.load(file) df_qc = pd.DataFrame([qc_data]) df_qc.columns = df_qc.columns.str.replace("_", " ", regex=False) #df_qc.columns = ['Value'] df_qc.style.hide(axis="index").format(precision=2) ``` # DNA over RNA counts Plotting normalized counts of DNA vs RNA (median across replicates). Only oligos within all replicates are shown. We should see a variation within the RNA count data (along the y axis). If count data between RNA and DNA is highly correlated (e.g. follows the identity line) there is no variation between designed oligos. This is an indication that RNA is inflated with DNA and the DNA digestion before creating cDNA did not work as expected. ::: {.panel-tabset} ## Threshold `{python} thresh` ```{python} #| echo: false Image(f"{workdir}/{dna_over_rna_thresh_plot}") ``` ## All ```{python} #| echo: false Image(f"{workdir}/{dna_over_rna_plot}") ``` ::: # Oligo correlation Oligo correlation plots of DNA, RNA and DNA/RNA ratios across replicates. First tab shows plots using (in average) `{python} "%d" % oligo_cor_merged_table_thresh['number_Oligos_Joined'].mean()` oligos with a minimum number of `{python} thresh` barcodes. Second tab shows all `{python} "%d" % oligo_cor_merged_table_all['number_Oligos_Joined'].mean()` oligos that have assigned barcodes. ```{python} #| echo: false def filter_table(df): df = df.round(2).drop(columns=['DNA_pearson','RNA_pearson', 'Ratio_pearson', 'fraction_Oligos_ReplicateA', 'fraction_Oligos_ReplicateB', 'NormSymmetry']) df.columns = df.columns.str.replace("_", " ", regex=False).str.replace("Replicate", " ", regex=False).str.replace("number ", "#", regex=False) return(df) ``` ::: {.panel-tabset} ## Threshold `{python} thresh` ```{python} #| echo: false oligo_cor_merged_table_thresh_filter = filter_table(oligo_cor_merged_table_thresh) oligo_cor_merged_table_thresh_filter.style.hide(axis="index").format( precision=2) ``` ## All ```{python} #| echo: false oligo_cor_merged_table_all_filter = filter_table(oligo_cor_merged_table_all) oligo_cor_merged_table_all_filter.style.hide(axis="index").format( precision=2) ``` ::: ::: {.panel-tabset} ## Threshold `{python} thresh` ::: {.panel-tabset} ### DNA ```{python} #| echo: false display(Image(f"{workdir}/{dna_oligo_coor_min_thre_plot }")) ``` ### RNA ```{python} #| echo: false display(Image(f"{workdir}/{rna_oligo_coor_min_thre_plot}")) ``` ### Ratio ```{python} #| echo: false display(Image(f"{workdir}/{ratio_oligo_min_thre_plot}")) ``` ::: ## All ::: {.panel-tabset} ### DNA ```{python} #| echo: false display(Image(f"{workdir}/{dna_oligo_coor_plot}")) ``` ### RNA ```{python} #| echo: false display(Image(f"{workdir}/{rna_oligo_coor_plot}")) ``` ### Ratio ```{python} #| echo: false display(Image(f"{workdir}/{ratio_oligo_coor_plot}")) ``` ::: ::: # Experiment statistic ```{python} #| echo: false df_merged = pd.read_csv(f"{workdir}/{statistics_all_merged}", sep='\t').round(2) n_oligos = df_merged['oligos design'].iloc[0] n_barcodes = df_merged['barcodes design'].iloc[0] avg_oligos = df_merged['oligos dna/rna'].mean() avg_matched_barcodes = df_merged['matched barcodes'].mean() avg_unknown_barcodes = df_merged['unknown barcodes dna/rna'].mean() ``` The total number of oligos in this experiment is `{python} f"{n_oligos}"` (defined by the assignment) with `{python} f"{n_barcodes}"` associated barcodes. In average across replicates we see `{python} f"{avg_oligos:.0f}"` from `{python} f"{avg_matched_barcodes:.0f}"` average barcodes in the count data and around `{python} f"{avg_unknown_barcodes:.0f}"` barcodes where not in the assignment. ::: {.panel-tabset} # DNA/RNA merged ```{python} #| echo: false df_merged.drop(columns=['oligos design','barcodes design', 'barcodes dna/rna'], inplace=True) df_merged.style.hide(axis="index").format( precision=2) ``` # DNA/RNA individual ```{python} #| echo: false df_single = pd.read_csv(f"{workdir}/{statistics_all_single}", sep='\t').round(2) df_single.style.hide(axis="index").format( precision=2) ``` ::: # Histograms barcodes per oligo, counts per barcode Histogramm of number of barcodes per oligo and the number of counts per barcode devidied by DNA and RNA. Median is red, mean is blue. ::: {.panel-tabset} # Number of barcodes per oligo ```{python} #| echo: false display(Image(f"{workdir}/{counts_per_oligo}")) ``` # Number of counts per DNA barcode ```{python} #| echo: false display(Image(f"{workdir}/{counts_per_bc_dna}")) ``` # Number of counts per RNA barcode ```{python} #| echo: false display(Image(f"{workdir}/{counts_per_bc_rna}")) ``` ::: # Activity Violin and box plots of the log2 fold change of all oligos by the assay. Grouped by labels if set, otherwise `NA`. First tab shows plots using (in average) `{python} "%d" % oligo_cor_merged_table_thresh['number_Oligos_Joined'].mean()` oligos with a minimum number of `{python} thresh` barcodes. Second tab shows all `{python} "%d" % oligo_cor_merged_table_all['number_Oligos_Joined'].mean()` oligos that have assigned barcodes. ::: {.panel-tabset} ## Threshold `{python} thresh` ```{python} #| echo: false display(Image(f"{workdir}/{activity_thresh}")) ``` ## All ```{python} #| echo: false display(Image(f"{workdir}/{activity_all}")) ``` :::